Temporal Context Discovery for User Device Content Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current recommendation systems fail to effectively incorporate temporal context to improve item recommendations, relying on manual time splits or surface parameters, and lack the ability to automatically discover contextualized recommendations.

Innovation Solution

A method that maintains records of user device consumption data with timestamps, identifies temporal consumption periods, clusters similar patterns, and creates user profiles based on recurring behaviors to provide time-specific recommendations by calculating weighting factors and decaying values, allowing for automatic discovery of temporal contexts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If manual time splits are used to incorporate temporal context, then temporal context can be provided, but the system complexity increases and automation is reduced

Engineering Contradiction:

Improveautomatic discovery of temporal contextsVSAvoidsystem complexity

Core Design Contradiction:

Extent of automationVSDevice complexity

Solution Approach 1:

The system automatically discovers temporal contexts by analyzing consumption patterns without requiring manual configuration. The algorithm self-adjusts to identify meaningful time periods based on user behavior data, eliminating the need for manual time split definitions while reducing operational complexity.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system dynamically adjusts temporal parameters by analyzing consumption data to automatically determine optimal time periods. Instead of using fixed manual time splits, the system adapts temporal context parameters based on observed user behavior patterns, improving automation while managing complexity through data-driven parameter adjustment.

Inventive Principle:

Principle #35Parameter changes

2Measurement precision

If surface parameters such as time and date are used to discover contexts, then automatic context discovery is enabled, but the recommendation accuracy is insufficient

Engineering Contradiction:

Improverecommendation accuracyVSAvoidautomatic context discovery

Core Design Contradiction:

Measurement precisionVSExtent of automation

Solution Approach 1:

The system uses feedback from consumption data to refine temporal context identification. By continuously analyzing user consumption patterns and comparing them against proposed temporal contexts, the system iteratively improves recommendation accuracy while maintaining automatic operation through data-driven feedback loops.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system replaces simple mechanical time-based filtering with a sophisticated data-driven approach that analyzes consumption patterns. Instead of relying solely on surface parameters like time and date, the system substitutes this with pattern recognition algorithms that process consumption data to identify meaningful temporal contexts, improving recommendation accuracy while preserving automation.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Quantity of substance

If older ratings are given more weight, then historical data is utilized, but the relevance of recommendations to current user preferences decreases

Engineering Contradiction:

Improveamount of historical dataVSAvoidrelevance of recommendations

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The system dynamically adjusts the weight of historical ratings based on temporal context. Instead of using static weighting, the system adapts the importance of older consumption data according to the identified temporal patterns, allowing historical data to contribute to recommendations while appropriately reducing its weight over time to maintain relevance with current user preferences.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements periodic analysis of consumption patterns to determine optimal weighting for historical data. By analyzing consumption data in identified temporal periods, the system periodically adjusts the weight given to older ratings, balancing the utilization of historical data with the need for current preference relevance through rhythmic, periodic evaluation and adjustment.

Inventive Principle:

Principle #19Periodic action

Data Source

PatentUS9330162B2Method and apparatus for providing temporal context for recommending content for consumption by a user device

Publication Date: 2016.05.03 TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

AI summary

A method for operating a system to provide temporal context for recommending items for consumption by a user device is described. The method comprises maintaining a record of items consumed by the user device or a group of devices within a reference period, together with the time of consumption of each item, and a content descriptor associated with each item. Temporal consumption periods are identified within the reference period, each consumption period spanning the consumption of one or more items with similar content descriptors, and each consumption period is associated with its respective content descriptor. An aggregated list is created of consumption periods recorded over a plurality of reference periods. Clusters of similar consumption periods are identified in the aggregated list, and recurring temporal patterns for user device behavior are identified in each cluster. A profile is created for each user device based on the clusters and the recurring temporal patterns for each cluster, and this profile is used to provide the temporal context at the user device.